IcemanG17

I think the stock non euro 16V's slow around 200CFM.....maybe less

Great results Greg!! Impressive flow numbers for the 16V head.....I think Deveks "Big Valve" heads flowed around 330CFM....with twice the valves!

slate blue

O.K guys the baseline flow number are here and they were in an earlier graph in the thread, so that is where the guys at the head shop left the job. I then took over and with their help with advice we kept going.



IN STD is the stock valve and port, it doesn't flow anymore after 0.600" which is the reason it stops there. The port was cleaned and the valve polished up so it wasn't a dodgy base line it was like new when tested.

IN PORTED 2 was quite a large port, around 2.5 sq" and you can see it is not as good as the smaller final port which was 2.35 sq"

The exhaust was also improved over these numbers and I recall peak numbers (at peak cam lift) in the 255 cfm range, again very respectable.


Taking the numbers from the graph as follows;

__Ported 2________Final Port
0.050" 30_________32.4
0.100" 61_________65.9
0.200" 130________137.4
0.300" 190________205.9
0.400" 236________254.7
0.500" 274________292.1
0.600" 298________311.4
0.700" ___________319.6

Now if you look at IN PORTED 1 there is points where it is better than IN PORTED 2 different shapes equal different efficiencies. My smaller port should like higher air speeds because it is quite straight and doesn't have a lot of choke point or major choke points. It doesn't look like the factory port although yesterday one port I tried was an enlarged factory port.

The numbers for that were

0.050" 33.6
0.100" 67.0
0.200" 137.4
0.300" 191.2
0.400" 235.6
0.500" 271.9
0.600" 300.1
0.700" 316.9
0.800" 325.6
0.900" 336.4
1.000" 340.2

Now you can see how these numbers just keep going up and you end with a higher number at 1.000" than you do with the final port but the area under the curve or cam profile is less. (up to 0.650") I also didn't really like the look of the factory port in terms of speed. So in the end it has worked out well. The one thing I will say in regards to the factory port is that it achieved a very good average flow per sq inch of valve area.

It achieved 86 cfm per sq", my design rounded did 85 cfm per sq". So not as efficient. The reason for this maybe a couple of things. I have a much bigger valve in proportion to bore area. Standard is 46% and mine is 52% so I have more shrouding. Two, the beginning of the port in the standard head is 40.5 mm to a 45 mm valve and mine is 43.5 mm to a 53.34 mm valve so in proportion much smaller as such the air will run faster in the runner in my head due to higher flow numbers. I am not going to lose sleep over it.

Brian the Devek or Bob DeVore head was quoted on the Rennlist as having a 2.18" valve and achieving flow at 0.500" of 300 to 305 cfm. I can not use a valve that large, my bore is 4.060". However I am making it back up by having higher lift cams, the DeVore motor quoted at 0.542" so about the same flow and same capacity, just different routes. Obviously that motor did a big number and nothing I have seen would contradict that potential. If I had I would have stopped. It has been a lot of work.

Greg

slate blue

In relation to what head was used it was an early euro head, however I didn't take that much material out and as such any head may be a candidate for this port design/program. As long as you don't get water. You can use an offset in the CNC program to make the cutting larger of smaller proportionately. You don't need to run a 2.10" valve.

I would agree Brian that the non euro heads would be less than 200 cfm. However you need to remember that your engine needs to demand more than that for that to be a problem. Airspeed will be a bigger problem in most cases. A lot of benefit can come from a tidy up. My first engine proved that, it was essentially stock but had been optimized with albeit an elaborate exhaust but that thing can really go.

Greg

danglerb

Will there be a way for others to use/tweak this same cnc code on other heads?

slate blue

O.K searched high and low and found the exhaust numbers, so I will try to compare them to the stock and first modified numbers.

Lift____Stock____Modified____Final___Percentage Improvement over stock
0.050"__23.1____24.9_______23.9____3.5%
0.100"__48.9____55.8_______58.5____20%
0.200"__107____114________121.4___13%
0.300"__135.4___147.8______167.6___24%
0.400"__162_____183_______194.8___20%
0.500"__172.6___194.4______196.1___14%
0.600"__172.6___213________218.8___26%
0.700"____________________225.3
0.800"____________________230.6
0.900"____________________234.4
@0.500 with the test pipe____242.4



Now exhausts are normally tested with a flow pipe, the flow pipe is a 90 degree pipe that should be the same diameter as the exhaust pipe. I remember at the time comparing the results with the AFR heads and thought I am dead level with most of their big small block heads on the intake and better on the exhaust side. So you can add 45 cfm at 0.500" and relative amounts at other lifts. So the relative increase would be bigger in reality but you need to keep consistency when you test anything.

So compared to the intake what you need to consider is that the exhaust valve didn't increase in anywhere near the same percentage as the intake. A 1.65" or 41.91 mm valve versus 40 mm, so about 10% bigger versus over 40% in surface area for the intake. So to give another comparison, against the AFR heads they did 221 cfm at 0.500 so around 20 cfm less than the Porsche head. So I hope this illustrates that these 30 year old castings can preform as well if not better than current CNC aftermarket heads.

By danglerb

Yes I believe so, maybe Mike Simard can chime in here. My knowledge of CNC is at the beginner stage but what I believe can be done is you set an "offset" and that offset effectively changes the cutting as a percentage but terms of the design, it is identical. The one proviso if that is correct you will need for the cutting tool to touch all parts of the port so that it is properly reshaped. That is one reason I believe the best way to go is the US heads as the port is smaller. I think you could weld the chamber up and then CNC the chamber also as that has been changed to achieve these figures.

I also think to get proper benefit you need as large a capacity as possible, 5.0 litres would be minimum. Airspeed will just be too low and that is because demand will be too low. Probably use a 49 mm or 48 mm intake valve and the 40 mm factory exhaust. Both 48 and 49 mm valves are available for the 928 off the shelf and they are quite cheap too. Well at least the 49 mm one is, around $20 from Pelican.

slate blue

I just got some good info i thought I would share, it will apply to all engine both 2V and 4V. I am trying to accurately size the intake port before CNC. The best flowing port has been posted and is in the above post. However it may be too small for the intended rpm range of the engine. We have discussed this a bit previously but this formula I have is for a live engine and not the normal 28" of depression most flow benches use. It may also give some insight into why certain engines preform very nicely and others not so good.

The PM I received is below.

Hi Greg,

Looking at the data I've collected over the years, all of the better engines I've measured inlet port CSA and were then dynoed seem to have from 600-650FPS at peak power rpm at the minimum CSA. That is using the formula-

FPS=(Bore^2 x stroke x RPM (At peak HP) x 0.00353)/Min CSA

All measurements in inches.

I know that still gives a fairly wide range of min CSA but at least gives you an idea. One engine actually had a calculated velocity of 783FPS. It was a 427 SBC, way too small 18 degree heads with a pushrod pinch of only 2.30sq". It had quite a big cam of high 270's on inlet, high CR and running exotic fuel. It still made about 720HP at 7500RPM. Inlet flowed 313cfm at 0.700" lift. With a better head it would have made closer to 800HP. However, the large cam made it peak artificially high I believe, much higher than it's "natural" peak for want of a better description, eg head really wants to peak somewhere around 6000rpm due to velocity getting too high but cam keeps engine going much higher. Plug the numbers into Pipemax. A 427ci making 718HP at 7500rpm is quite poor VE. This is what a mismatch of parts generally does. Although it did demonstrate what you can do if you had CSA restrictions on your head casting but wanted more power. This 18 degree head though could have been easily improved.

One other point to keep in mind is the expansion rate from min CSA to throat. You have quite a short inlet port so it may help to err on the larger side to keep this expansion more gentle. Like any venturi, it's what happens after the choke that's more important.

That Windsor I was telling you about with the AFR205cc CNC heads has a 4.125" bore and 3.40" stroke, about 363ci.
14" long IR EFI with 4.250sq" CSA at runner entrance. Those heads have a min CSA at pushrod of 2.20sq". From memory they have a 2.08" inlet valve. Made approx 660Hp at 7000rpm on Kim Bakers dyno last year. If you recall our discussion last month, the IR manifold was also incorrectly designed which dropped about 20-30cfm at peak as it was angled the opposite way to the port.





The port is made bigger because of my concerns with a different formula, that formula was the sizing of the port using the average CSA with the flow at 28" of depression on the flow bench. So what that means in effect is that if you say have 300 cfm of flow available you need to size the port at a certain size to achieve 311 FPS (28" formula) at that 300 cfm flow rate. So you could have a too big port why by the port speed is too slow to achieve that 311 FPS or again like my port the port is too fast. I asked my friend that question and to please give his opinion.

Now the other thing to remember is that the engine must be able to create that demand for the air, if it can't the air speed again will be too slow and you wont get your full ram effect. One of those formulas is as follows, CFM =311xCSA/2.4

So my problem was this, peak flow is 315 cfm. Therefor 311 x 2.25 sq" / 2.4= around 295 cfm so I was losing a 20 cfm potential or around 50 hp just in sizing. So I did make the port bigger and lost some flow in the curve but not peak flow but hope to recover this later. So now I have 311 x 2.45 sq" divide by 2.4 = 317 or the full potential of the port.

So the S4 port is 2.42 sq" when I measured it and using the formula for a live engine above gives the following numbers, 3.937 x 3.937 x 3.11 x 6000 x 0.00353 = 1020 divide by the CSA of 2.42 and you get 421 Feet per second or roughly two thirds of the optimum speed.

A stroker is different, using 4.1 bore and 3.75 stroke and 6,500 rpm the sum is 1446 divided by 2.42 sq = 597 FPS much closer to that optimum range and the 0.55 MACH number that all the engineers aim for which is around 615 FPS. Of course use a few more revs and you will be at that peak number. Also it could explain the very good boosted engine results too. Of course if you increase the port size you will need to increase the engine's demand for air through size or revs.

What I have also had to do with the 2V port to try and keep the CSA as even as I can throughout its length, not easy and it will require a fair bit of machining for this bigger engine. (not required on smaller engines) I plan to make a plate to go between the head and the cam box of around 10 mm thickness. I will weld up the spring seats and have them recut. The springs will be able to be longer because of this mod and as such be able to cope easier with the increased lift.

The lift of 0.650" and the cam bucket diameter is basically what the Pro Stock motorcycle guys ran before radiused buckets cam in. They make for the sake of conversation 3.4 hp per CI at 14,000 rpm.

Anyway hope that is of interest.

Greg

Jim Devine

Weaver Bros make great pumps! They have been doing it since the 1960's & have extensive knowledge of what really works. Their products
have been used by most major teams in all forms of racing. Well worth talking to.

http://www.weaverbrothers.com/

Good luck with the project!

Jim Devine

Try Kinsler- again, like Weaver dry sump pumps, Kinsler fuel injection parts are the highest quality.
Have a look at this page & contact them for your needs.

http://www.kinsler.com/page--Throttle-Bodies--22.html

beentherebaby

Nice project Greg.

I would not aspire to have a peak port velocity above Mach .5 based on my engine design and development work. You may desire a peak velocity of ~550 FPS using the minimum port CSA. If you can incorporate a 2-5% taper on the intake runners that is useful. Exhaust flow typically only needs to be 65% of a good intake port design. Many American V8 engines have a localized port restriction, i.e. reduced CSA. Using that localized restriction as the peak velocity reference point can be misleading and confusing compared to a proper port design of relatively uniform CSA.

slate blue

By Beentherebaby

Basically I agree with everything you say in general terms. As far as taper goes I have been told to use no more than 2 degrees per side, so your number may be inclusive. My exhaust cam lobe will be smaller than the intake, unlike some american V8s like the ford cleveland.

Now all you say may be 100% accurate when applied to different engines, example may be an early 23 degree chevy head, they turn quite abruptly and as such may develop a localized choke point due to excessive air speed. In that case the average air speed may be deceptive, which I think is your point?

In my case though the port is very straight and uniform. I am hoping that it has uniform air speed too. What I had to do and this is not all bad. Is widen the port at the start of the short side, this normally has the effect to slow the air down in that critical area as it is normally the area where the air goes way too fast. Also I raised the roof, now that makes the roof very thin, that is why welding up the spring seat and cutting a new one higher up, say 4 mm will hopefully allow me to run sufficient spring pressure and have another 6 mm for bigger longer spring more capable of 0.650" lift. Raising the roof also straightens the port.

I can't slow the air anymore, if I do, I can't rev the engine to the limits it has been designed for. The limit will be 8,500 but wait till we see the power and that may be revised. No revving it for the sake of it.

Greg

slate blue

By Jim devine

hi Jim, I have sorted a number of parts suppliers, I am going to use Auto Verdi, there will be some pumps coming up soon, it will be a used pump from Nascar, Gibbs, Yates or some other engine shop, I couldn't afford to buy a new Verdi, they are the Rolls Royce of pumps, super light too just 4 kgs complete, around $4K new. They pull a huge vacuum with their roots lobes and I will use their oil pump and delete the Porsche one.

The Kinsler stuff I am aware of, have been a fan since the 70s, however it is orientated to american pushrod V8s. I am hopefully going to be able to convince these fellows in England to make throttles for the 928 and 944. http://www.atpower.co.uk/main.html They make the most advanced throttle bodies available today, no shafts! With the way the pound is there is a great opportunity there along with the engine management from Cosworth, the SQ6M.

Greg

Jim Devine

Great parts! Weaver & Kinsler parts are excellent,
but you are going to the next level!

Found this: Not a current ad, not sure if it's still
available, but the contact info is there if you want to try:

http://limerock.racingjunk.com/categ...-Pump-New.html

Best of luck with your engine!

slate blue

OK I have been dealing with this port again, now the existing port will run exceptionally well, have no doubt in that regard. Notice how the new designs straighten the port. The port on the paper is Darin Morgan's port. That port is for a Pro Stock Suzuki which makes 3.4 hp per CI. I didn't copy that port it is just coincidental. The port in that pic that looks like doggy do do was tested and is the best at the top end, 286 cfm ay 0.500" and 325 cfm at 0.650" or peak cam lift. The important 87% lift number is 306 cfm, that area is often where peak piston demand occurs.



So I made new molds, the best port so far at 0.500" is the middle mold, the one on the left is the stock port.






The port on the right is the untested port. It has been designed to feed the engine for peak power at 8,000 RPM. It has a very consistent CSA and the short side is as well preserved as I can do. This head will require a 10 mm spacer between the head and cambox.

The steeper the short side the better the flow will be in the lower lift ranges, given 928 cams tend to be low lift cams it is obviously an important consideration.

I honestly have no idea what the larger port will flow, it is around 5% bigger. I am hoping for an increase from 0.500" upwards and a big increase from 0.600" 335 to 340 cfm. That puts the engine in the low to mid 700 hp range. The existing port is in the low 700 hp range and the reason I mention this is that the Devek motor that over revved on the dyno, to around 7,800 rpm and was 6.2 litres in capacity had similar flow figures from my research and we are running into the same problems like hitting water.

Anyway I am sure you can see how much straighter the ports are and that is the main reason for the increased efficiency. Air never likes sharp turns hence the relatively poor performance of the S4 manifold in performance applications. Keep it straight!

Greg

GregBBRD

Very nice work. You have air flow that is getting very close to trick Big Block Chevy heads with huge valves...that is difficult to get one's "head" around. (Sorry for the pun.)

Your port shape is fairly straighforward. Nice work around the valve seat. Regarding your "middle" port. Here's something for you to ponder....If you look at your port shape from side view (yeah, I know, both are side views), there isn't much "pinch" 1/2" from the seat. In the other view, there is considerable "pinch" just before the seat. It would be interesting to cross section your mold and actually measure this area, to see if either reducing the "side" pinch might help...

Keep up the good work!

beentherebaby

I can easily see why the center port flows well at .500" lift.

A couple points worth noting, the valve is only at max lift for a very short real time/duration compared to the durations at less than max lift.

Darin Morgan is a smart guy but following what works on a Chevy/Ford/Mopar/Suzuki motorcycle, (think weight per C.I.), for a specific race type application can lead you far, far away from where you would ultimately prefer to be performance wise.